Dell Storage Solution Resources Owner's manual

Brand: Dell

Model: Storage Solution Resources

Type: Owner's manual

Dell Storage Solution Resources provides state-of-the-art data storage solutions tailored for businesses of all sizes. Its robust capabilities empower users to safeguard valuable data, maximize storage efficiency, and optimize performance.

With Dell Storage Solution Resources, you can:

Centralize data storage: Consolidate data from various sources onto a single, unified platform, simplifying data management and enhancing accessibility.
Ensure data protection: Implement comprehensive data protection strategies with features like snapshots, replication, and backup, ensuring business continuity and minimizing data loss risks.

With Dell Storage Solution Resources, you can:

Centralize data storage: Consolidate data from various sources onto a single, unified platform, simplifying data management and enhancing accessibility.
Ensure data protection: Implement comprehensive data protection strategies with features like snapshots, replication, and backup, ensuring business continuity and minimizing data loss risks.

A Dell EMC Best Practice

Dell PS Series Storage Arrays iSCSI Initiator

and Operating System Considerations

Optimizing Your SAN Environment for High Availability

Dell Storage Engineering

January 2017

2 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

Revisions

Date

Description

October 2013

Initial release

January 2017

New format, Added XenServer, Windows Server 2012 R2

The information in this publication is provided “as is.” Dell Inc. makes no representations or warranties of any kind with respect to the information in this

publication, and specifically disclaims implied warranties of merchantability or fitness for a particular purpose.

Use, copying, and distribution of any software described in this publication requires an applicable software license.

Other trademarks may be the property of their respective owners. Published in the USA [1/12/2017] [Best Practice] [110-6165]

Dell EMC believes the information in this document is accurate as of its publication date. The information is subject to change without notice.

3 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

Table of contents

Revisions............................................................................................................................................................................. 2

Preface ................................................................................................................................................................................ 5

Intended Audience .............................................................................................................................................................. 5

1 High-Availability Procedure Summary .......................................................................................................................... 6

1.1 Proactive Maintenance Support ......................................................................................................................... 6

1.2 Host Integration Tools ........................................................................................................................................ 6

2 About High-Availability iSCSI Data Transfer ................................................................................................................ 7

2.1 Before You Begin: Update Initiators and Drivers ................................................................................................ 7

3 Configure Your Hosts for High Availability ................................................................................................................... 8

3.1 Host Integration Tools ........................................................................................................................................ 8

3.2 Operating System Recommendations ................................................................................................................ 8

3.2.1 Test Modifications in a Non-Production Environment ........................................................................................ 8

3.2.2 Increase Transmit and Receive Resource Settings ........................................................................................... 8

3.2.3 Operating System Specific Configuration Information........................................................................................ 8

4 Configure Your Network for High Availability ............................................................................................................... 9

4.1 General Network Requirements ......................................................................................................................... 9

4.1.1 Design Network for Redundancy ........................................................................................................................ 9

4.1.2 Ports ................................................................................................................................................................... 9

4.1.3 Switches ............................................................................................................................................................. 9

4.1.4 FS Series Array Switch Requirements ............................................................................................................. 11

4.2 Hardware Initiators............................................................................................................................................ 11

A Configure Your Microsoft Windows Operating System .............................................................................................. 12

A.1 Procedure Summary ......................................................................................................................................... 12

A.1.1 Install the Recommended Microsoft Hotfixes ................................................................................................... 12

A.1.2 Install the Dell HIT/Microsoft tools. ................................................................................................................... 14

B Install the Multi-extension module for VMware operating system .............................................................................. 15

C Configure your Linux operating system ...................................................................................................................... 16

C.1 Configure SCSI Device Settings Using the udev Facility ................................................................................ 16

C.2 Configure iSCSI Settings Using the /etc/iscsi/iscsid.conf File .......................................................... 16

C.3 Configure Linux Networking ............................................................................................................................. 17

D Configure your Oracle Solaris 10 operating system................................................................................................... 19

D.1 Zero-Length Volumes ....................................................................................................................................... 19

4 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

E Configure your IBM AIX operating system ................................................................................................................. 20

E.1 iSCSI Initiators Lose Volume Connectivity ....................................................................................................... 20

F Configure Your Citrix XenServer Operating System .................................................................................................. 21

G Technical support and resources ............................................................................................................................... 22

G.1 Related resources............................................................................................................................................. 22

5 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

Preface

Dell™ PS Series arrays optimize resources by automating performance and network load balancing.

Additionally, PS Series arrays offer all-inclusive array management software, host software, and free firmware

updates. Dell FS Series appliances, combined with PS Series arrays, offer a high-performance, high-

availability, scalable NAS solution.

This document describes configuration information, limitations, and other considerations required to architect

a PS Series SAN for high availability and preparation for non-disruptive array firmware upgrades.

Intended Audience

The information in this guide is intended for experienced administrators who are responsible for managing a

PS Series group. To use the info1mation in this manual, you must have an understanding of the following

topics:

 PS Series group administration

 Operating environment administration for:

o - Windows

(or Windows Server) systems

o - VMware

o - Linux

systems

 Basic network and SAN administration

6 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

1 High-Availability Procedure Summary

This chapter summarizes the procedures that you must perform to optimize your environment for high

availability (HA).

1.1 Proactive Maintenance Support

If you have a Dell proactive service agreement, contact your Dell Technical Support representative, who will

help you analyze your SAN environment and ensure it is correctly set up for high-availability.

1.2 Host Integration Tools

Dell has automated a portion of this process in the PS Series Host Integration Tools (HIT) kit. If installed, this

software implements changes such as registry settings and configuration options on the hosts that follow

Dell's recommended best practices. HIT is available for Microsoft Windows, Linux and VMWare. Downloads

and more information is available at https://eqlsupport.dell.com (requires login).

Read the following sections in this document and perform the recommended actions:

 About High-Availability iSCSI Data Transfer - provides an overview of high-availability iSCSI

data transfer.

 Configure Your Hosts for High Availability - contains general operating system

recommendations and guidelines to configure settings on your servers to ensure high-

availability.

 Configure Your Network for High Availability - outlines the general best practices required to

configure a redundant iSCSI network.

7 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

2 About High-Availability iSCSI Data Transfer

This chapter describes how to optimize your storage area network (SAN) environment for highly available

(HA) iSCSI data transfer. It contains information about how to:

 Configure Your Hosts for High Availability

 Configure Your Network for High Availability

Note: If you have a Dell proactive service agreement, your Dell Technical Support representative can help

you analyze your SAN environment and ensure it is correctly set up for high availability.

For more information about Proactive Maintenance, visit dell.com/services.

When you optimize your SAN environment for HA, you increase the fault tolerance of the SAN, providing

protection from component failures and minimizing service disruptions.

Many causes of service disruptions can occur in a data center. Some of the more common ones are:

 Hardware failure in servers, NICs, cables, switches, storage controllers, and disks

 Software failure such as bugs or application crashes

 Upgrades to software, drivers, or hardware firmware

In most cases, if your da ta center is correctly designed for redundancy, the SAN can tolerate service

disruptions. In some cases, however, a redundant design is not sufficient to allow applications and operating

systems to continue to operate during disruptions. For example, problems with software or updates to

software and firmware can cause disruptions that redundancy cannot manage. In these cases, other best

practices are typically required to tune applications and opera ting systems to continue to operate without

being affected. This document describes the actions you can take to ensure a highly available iSCSI data

transfer that minimizes the possibility of these issues with PS Series SANs.

2.1 Before You Begin: Update Initiators and Drivers

Dell recommends that you keep initiators and drivers up to date. A list of supported initiators can be found at

http://en.community.dell.com/techcenter/storage/w/wiki/2639.ps-series-configuration-guide. Apply the

recommended modifications to your operating system and iSCSI initiators specified in the following sections.

The following chapters contain configuration and performance tuning best practice recommendations for high

availability and non-disruptive operation with PS Series SANs.

8 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

3 Configure Your Hosts for High Availability

3.1 Host Integration Tools

Dell has automated a portion of this process in the PS Series Host Integration Tools (HIT) kit. The PS Series

HIT kits provide integrated software such as MPIO modules, enhanced data protection, automated

provisioning, and alerts for Windows, VMware, and Linux hosts. If installed, this software implements changes

(such as configuration options on the hosts) that follow Dell's recommended best practices.

You can download the PS Series Host Integration Tools (HIT) kit from the https://eqlsupport.dell.com (login

required) website, and install it on supported operating systems.

3.2 Operating System Recommendations

This section contains detailed information about settings and configuration options you should apply to hosts

running in your PS Series iSCSI SAN to ensure high availability (HA) at the host level.

3.2.1 Test Modifications in a Non-Production Environment

If you are not running any of the operating systems or iSCSI initiators listed here, or if you are unsure of your

applications’ HA tolerance, Dell recommends that you first test the suggested modifications in a non-

production environment.

3.2.2 Increase Transmit and Receive Resource Settings

A PS Series group might deliver more network I/O than iSCSI initiators can handle, resulting in dropped

packets and retransmissions. Some iSCSI initiators, however, let you increase the transmit and receive

resource settings, which can help minimize this problem.

For example, on a Windows system, you might be able to increase the initiator's network interface settings for

receive and transmit descriptors. On a UNIX or Linux system, increase the transmit and receive resources for

your network ports.

3.2.3 Operating System Specific Configuration Information

To view configuration information specific to your operating system, refer to the following appendixes:

A: Configure Your Microsoft Windows Operating System

B: Configure Your VMware Operating System

C: Configure Your Linux Operating System

D: Configure Your Oracle® Solaris 10 Operating System

E: Configure Your IBM® AIX® Operating System

F: Configure Your XenServer Operating System

9 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

4 Configure Your Network for High Availability

A PS Series iSCSI SAN can be operated in any network that supports the industry standards and IP subnet

design guidelines described in this chapter. Because of this flexibility, many network design and configuration

choices can affect SAN performance.

For more network design and configuration information to support the use of a PS Series SAN, see the Dell

PS Series Configuration Guide, which can be downloaded from:

http://en.community.dell.com/techcenter/storage/w/wiki/2639.ps-series-configuration-guide

4.1 General Network Requirements

This section outlines the general best practices for configuring a redundant iSCSI network.

4.1.1 Design Network for Redundancy

For PS Series PS arrays, follow these general SAN design requirements:

4.1.2 Ports

The arrays should be in the same subnet as the group’s IP address. For all members (arrays) in a given SAN

group, all ports should be connected to the same subnet. This connection allows the arrays to communicate

with each other as a group of peer members.

At least two iSCSI SAN ports per host must be connected to two different switches in a SAN fabric for fully

redundant SAN connectivity.

4.1.3 Switches

All switches within the SAN must be interconnected such that a path is always present from any

Ethernet port on one array to all other Ethernet ports on all other arrays in the group.

Enabling jumbo frames on all switches in the SAN is a recommended best practice as long as the rest of the

infrastructure (such as initiators and NICs) supports jumbo frames as well.

Rapid Spanning Tree Protocol must be enabled if the SAN infrastructure has more than two switches in a

non-stacked configuration, and portfast must be enabled on all edge device ports (hosts, and FS Series

appliances and arrays).

To prevent a switch failure from also disabling all paths between a host and its connected volumes, all ports

from each controller must be connected to at least two different switches. This requirement also applies to FS

Series NAS appliance configurations.

For PS4100/PS4110/PS4210/PS6100/PS6210/PS6610 family arrays, split the vertical port pair connections

between two switches to ensure 100% bandwidth capability is maintained in the event of a vertical port

failover event, as shown in Figure 1 and Figure 2.

10 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

All switches in the SAN must have jumbo frames enabled. This requirement applies to SANs connected to

either a PS Series array or an FS Series appliance.

Redundant SAN PS4100

Redundant SAN PS6100

11 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

 Both LAG or Stacking interconnects can be used as the bridging mechanism between

redundant switches in the PS Series SAN, but each method has a different set of

considerations that might make one approach more preferable than the other. For more

information about LAG and Stacking considerations, see the Dell PS Series Configuration

Guide at: en.community.dell.cornltechcenter/storage/w/wiki/2639.equallogic-configuration-

guide.aspx

 Any PS Series SAN group that is required to send or receive replication traffic to or from

another SAN group must have an uninterrupted communications path (such as "visibility")

between each group.

See Dell's Switch Configuration Guides for specific switch vendor and model settings and best practices.

Examples and detailed instructions about configuring SAN network switches and hosts can be obtained

through the Rapid PS Series Configuration Portal or by reviewing sample SAN reference architectures on the

Dell TechCenter website: en.community.dell.cornltechcenter/storage/w/wiki/4250.switch-configuration-

guides-by-sis.aspx

4.1.4 FS Series Array Switch Requirements

 Keep the client and SAN side networks physically separate, and deploy two switches on each

side to protect against a switch failure.

 Do not block IPv6 traffic on your SAN/internal switches if you have FS series appliances

attached.

4.2 Hardware Initiators

Most hardware initiators are referred to as Converged Network Adapters (CNAs). CNAs support a variety of

enhancements such as iSCSI Offload, DCB, and so on, making them more than just a hardware initiator. Of

the major CNAs supported by PS Series, the most common ones are QLogic®, Broadcom®, Emulex®, and

Intel®.

12 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

A Configure Your Microsoft Windows Operating System

The following sections describe how to configure settings on your Microsoft® Windows servers to ensure high

availability.

A.1 Procedure Summary

To ensure high availability, configure the settings on your Microsoft® Windows servers as follows. If you have

installed the HIT kit, skip step 3.

1. Install the recommended Microsoft hotfixes.

2. Install the HIT kit and proceed to step 4.

3. (Optional) Set the registry entries if the HIT kit is not installed.

4. Check the Microsoft guidelines.

A.1.1 Install the Recommended Microsoft Hotfixes

Note: Not all hotfixes are necessary for all versions of Windows. Dell recommends that you install the latest

Microsoft recommended hotfixes.

For Windows Server 2012, Windows Server 2012 R2 a few hotfixes are listed here as examples, please

check the Microsoft sites for the complete list of appropriate hotfixes.

 Microsoft has released a KB Rollup for Windows Server 2012 R2 MPIO LUN loss on storage

controller failover causing the Server to lose its Storage connection.

Assume that you have a Windows Server 2012 R2-based computer connected to a storage device

through multiple paths by using Multipath I/O (MPIO). When multiple paths fail over concurrently or in

quick succession, the I/O operation may fail, and the computer may lose its connection to the storage

device.

Recommendation: To resolve this issue, install the May 2016 update rollup for Windows RT 8.1,

Windows 8.1, and Windows Server 2012 R2 (KB3156418).

Related Link: (Microsoft Support Website – Article ID: 3156418

https://support.microsoft.com/en-us/kb/3156418

 Windows Server 2012 R2 connecting to an iSCSI SAN where the server unexpectedly

restarts on a high-speed network.

Note: To apply this update, you must first have update 2919355 installed in Windows 8.1 or Windows

Server 2012 R2. Recommendation: To resolve this issue, install the update rollup that is dated

November 2014. Microsoft has released hotfix KB2851144 to correct a LUNs tracking after a path

failover occurs between storage controllers in Windows Server 2008 R2 or Windows Server 2012.

 LUN not appearing in Disk Management after path failover.

You configure a Windows Server 2008 R2-based or Windows Server 2012-based computer to use

Microsoft Multipath I/O (MPIO) and Microsoft Device Specific Module (MDSM) or the Dell

13 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

HIT/Microsoft toolkit. A path failover occurs between storage controllers. In this scenario, the system

loses track of one or more logical unit numbers (LUNs). Therefore, not all the LUNs appear in the

Device Manager. Additionally, the system no longer sends out I/O requests to the missing LUNs.

Note: To resolve this issue in Windows Server 2012, install update rollup 2889784

Recommendation: Microsoft has released 2851144 hotfix to address this issue. If you have questions

regarding the 2851144 solution, please contact Microsoft Support.

Related Link: (Microsoft Support Website – Article ID: 2851144)

https://support.microsoft.com/en-us/kb/2851144

For Windows Server 2008 R2, visit the Microsoft website and then install the following hotfixes:

Note: The Dell HIT/Microsoft 4.8 and above does not support Windows Server 2008 or Windows Server

2008 R2.

 Update Windows Server 2008 R2 Clusters

If your environment contains Microsoft Windows Server 2008 R2 clusters, apply the hotfix specified in

Microsoft Knowledge Base article 978562, support.microsoft.com/kb/978562, to each cluster node

before upgrading group members.

 During a Failover or Load Balance Event, Volumes May Become Unavailable

If you are running multipath I/O in a Windows Server 2008 or Windows Server 2008 R2 environment

(might also affect clusters), and the server must disconnect and then reconnect to the volume (such

as for failover or load balance), the volume might become unavailable. If this situation occurs, refer to

Microsoft Knowledge Base article 2718576, support.microsoft.com/kb/2718576, for detailed

information on how to proceed.

 Some iSCSI Initiator Sessions Not Being Deleted

If you are running a large cluster configuration using Windows Server 2008 R2 without SP1 installed,

some iSCSI attempts to log out iSCSI sessions in the iSCSI initiator GUI might fail with an error

message indicating that the session cannot be logged out. In addition, if you are using the Host

Integration Toolkit, this can result in a higher than expected total iSCSI session count.

To fix the issue, install the hotfix specified in Microsoft Knowledge Base article 979710 to each cluster

node. See support.microsoft.com/kb/979710 for more information.

 Bugcheck 0x0000000A Returned After Control Module Failover

If you are running a large cluster configuration using Windows Server 2008 R2 without SP1 installed,

you might encounter bugcheck 0x0000000a after an array control module failover:

0x0000000A

(<parameter1>,<parameter2>,<parameter3>,<parameter4>IRQL_NOT_LESS_OR_EQUAL

14 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

To fix this problem, apply the hotfix specified in Microsoft Knowledge Base article 979711,

support.microsoft.com/kb/979711, to each cluster node.

 Performance Degradation When Heavy Disk I/O Load

The performance of a computer that is running Windows Server 2008 R2 or Windows 7 decreases by

20 percent under a heavy disk I/O load. Apply the hotfix specified in Microsoft Knowledge Base article

982383, support.microsoft.com/kb/982383.

 MPIO Driver Fails Over All Paths

The MPIO driver incorrectly fails over all paths when a transient single failure occurs in Windows

Server 2008 or in Windows Server 2008 R2. Apply the hotfix specified in Microsoft Knowledge Base

article 2522766, support.microsoft.com/kb/2522766.

 MPIO Does Not Fail Over Correctly

MPIO failover fails on a system that is running Windows Server 2008 R2 system and the path failover

process fails, any application that tries to access the storage device might stop responding.

Therefore, the MPIO path does not fail over. Apply the hotfix specified in Microsoft Knowledge Base

article 2460971, support.microsoft.com/kb/2460971.

 MPIO.sys Driver Error in Windows Server 2008 R2

If you connect a storage device that is controlled by MPIO to a Windows Server 2008 R2 system, and

if the system is under heavy backup workload, Windows Server 2008 R2 crashes. Apply the hotfix

specified in Microsoft Knowledge Base article 2511962.

A.1.2 Install the Dell HIT/Microsoft tools.

Dell has automated a portion of this process in the PS Series Host Integration Tools (HIT) kit. You can

download the PS Series HIT kit from the https://eqlsupport.dell.com website, and install it on supported

operating systems.

The PS Series HIT kit provides integrated software such as MPIO modules, enhanced data protection,

automated provisioning, and alerting. If installed, this software implements changes that follow Dell's

recommended best practices for the following registry values:

 Disk timeout

 EQL DSM SCSI timeout

 EQL DSM path verification enabled

 EQL DSM path verification period

 Link downtime

 Enable iSCSI NOP Out

 MPIO path recovery interval

 MSISCSI IO timeout

15 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

B Install the Multi-extension module for VMware operating

system

VMWare vSphere provides the Pluggable Storage Architecture (PSA), which enables multipath configuration

for high availability and increased performance. The modular design of the architecture accepts third-party

multipathing plugins that enable enhanced functionality specific to storage device models.

The Dell PS Series Multipathing Extension Module (MEM) provides the following enhancements to the

existing VMware multipathing functionality:

 Increased bandwidth

 Reduced network latency

 Automatic connection management

 Automatic load balancing across multiple active paths

For these reasons Dell highly recommends that the MEM be downloaded and installed on the VMware

servers. The MEM and other VMware integration tools such as Virtual Storage Manager for VMware, Storage

Replication Adapter for VMware Site Recovery Manager and Adapter for VMware vRealize Operations

Manager from https://eqlsupport.dell.com.

16 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

C Configure your Linux operating system

If you do not have the HIT kit installed, the following sections describe the configuration settings you need to

apply on your Linux hosts to ensure high availability. See Table 1 for a summary of the recommended

settings.

C.1 Configure SCSI Device Settings Using the udev Facility

The Linux udev utility can be used to dynamically modify device settings every time a new device is added or

changed. The following examples show ways that the HIT/Linux kit achieves the following settings with udev

rules:

 Increase SCSI Device Timeout Value to Avoid Disk I/O Timeouts

The default disk timeout should be increased to 60 seconds. The specific udtev rule implemented by

HIT/Linux is installed in the file named /lib/ucdev/rules.d/99-eqlsd.rules:

ACTION!="remove", SUBSYSTEM=="block", ENV{ID_VENDOR}=="EQLOGIC",

RUN+="/bin/sh -c 'echo 60 >/sys/%p/device/timeout'"

 Increase SCSI Device Read-Ahead Buffer Size

The default SCSI device read-ahead buffer setting of 128KB is increased to 1024KB. The specific

udev rule implemented by HIT/Linux is installed in the file named /lib/udev/rules.d/38-

equallogic.rules: ATTR{queue/read_ahead_kb}="l024"

 Change SCSI Device I/O Scheduler Algorithm

The default SCSI device I/0 scheduler algorithm of CFQ (Completely Fair Queuing) is set to noop.

The specific udev rule implemented by HIT/Linux is installed in the file named /lib/udev/rules.d/38-

equallogic.rules: ATTR{queue/scheduler}="noop"

C.2 Configure iSCSI Settings Using the /etc/iscsi/iscsid.conf

File

The following changes should be made in the Linux iSCSI configuration file /etc/iscsi/iscsid.conf. Any settings

in this file are propagated to iSCSI targets on discovery or rediscovery.

 Change FastAbort Setting in iSCSI Configuration File

The default iSCSI initiator setting for FastAbort is set to Yes. PS Series targets prefer this iSCSI

configuration parameter be set to No. The HIT/Linux software automatically implements this change:

node.session.iscsi.FastAbort = No

 Change node.startup Setting to Avoid Login Storms

The default iSCSI initiator setting for node.startup is set to automatic, which means that all previously

discovered iSCSI targets will be logged in to at system boot. This behavior can cause a surge of

iSCSI login activity during system reboots, which can place a resource burden on both the PS Series

17 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

arrays and the Linux host. Dell recommends that you change the setting to manual. The HIT/Linux

eqltune utility recommends this change, but does not automatically implement it:

node.startup = manual

C.3 Configure Linux Networking

Dell's best practice recommendation is to configure multipathing for Ethernet interfaces on a Linux host to

connect to a PS Series group. This multiple NIC per subnet configuration requires that several Linux

networking parameters be modified from default values for reliable implementation. The HIT/Linux eqltune

utility, when run in fix mode, automatically configures the Linux operating system to operate correctly with the

multiple NIC per subnet configuration.

Note: Apply these commands, changes, and settings to each Ethernet interface (for example: eth1, eth2).

 Disable Generic Receive Offload

The Generic Receive Offload (GRO) feature, implemented by some Linux network drivers, causes

poor read performance results. This feature can be disabled using the ethtool utility:

#ethtool ethX -K gro off

 Enable Flow Control

Flow control is often negotiated properly, but a network that supports full flow control (RX and TX

pause names on) is recommended. If your network infrastructure is set up to support flow control, you

could benefit from forcing flow control to be on. The HIT/Linux eqltune utility recommends this

change, but does not automatically implement it:

#ethtool –pause ethX autoneg off rx on tx on

To make this change persistent across reboots, add the command to the /etc/rc.local directory.

 Enable Jumbo Frames

If supported by your networking infrastructure, enabling jumbo frames (that is, an MTU of 9000)

enhances iSCSI throughput under heavy load. The HIT/Linux eqltune utility recommends this

change, but does not automatically implement it:

In /etc/sysconfig/networking/devices/ifcfg-ethX (where X is the Ethernet port) add this line:

MTU=9000

 Avoid ARP Flux

The default value of 0 (zero) allows NICs to reply to ARP requests on any other NIC, which can

provide false positives to externally originating ping tests, as well as unexpected ARP resets. This

condition is known as ARP Flux. To avoid ARP Flux in a multiple NIC per subnet configuration, apply

the following parameters in /etc/sysctl.conf:

net.ipv4.conf.ethX.arp_ignore = 1

net.ipv4.conf.ethX.arp_announce = 2

18 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

 Loosen IPv4 Reverse Path Filtering

To allow multiple iSCSI connections in a multiple NIC per subnet configuration, reverse path filtering

must be loosened. Use this setting:

net.ipv4.conf.ethX.rp_filter=2

Summary of Recommended Changes

Area

Setting

Default Value

Recommended Value

(PS Series)

SCSI block devices

Device timeout

30 seconds

60 seconds

Per SCSI device

read-ahead buffer

128KB

1024KB

Per SCSI device

I/O scheduler

CFQ

noop

iSCSI settings

node.session.iscsi.FastAbort

Yes

iSCSI settings

node.startup

manual

automatic

Ethernet devices

Generic Receive Offload

off

Ethernet devices

flow control

auto-negotiate

Ethernet devices

jumbo fiames

(MTU size)

1500

9000

Sysctl tunable - ARP

net.ipv4.conf.ethX.atp_ignore

Sysctl tunable - ARP

net.ipv4.conf.ethX.atp

announce

Sysctl tunable - ARP

net.ipv4.conf.ethX.atp filter

Verify the network infrastructure supports these two settings (flow control and jumbo frames) before

implementing them.

19 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

D Configure your Oracle Solaris 10 operating system

You can use the default initiator configuration for Oracle

Solaris

10 operating systems. No changes are

required. However, you might experience the following issue.

D.1 Zero-Length Volumes

By default, the Solaris 10 initiator automatically connects to all volumes it discovers. If it connects to a zero-

length volume (such as the vss-control volume), it panics. To avoid this issue, make sure the vss-control

volume on the group has access control records configured to prevent the Solaris initiator from discovering it.

In other words, make sure the volume does not allow unrestricted access.

20 Dell PS Series Storage Arrays iSCSI Initiator and Operating System Considerations | 110-6165

E Configure your IBM AIX operating system

The AIX native iSCSI controller does not perform device discovery, so you will need to know your iSCSI IQN

target names specifically. In addition IBM

AIX

systems might experience the following issue.

E.1 iSCSI Initiators Lose Volume Connectivity

In some cases, after a load-balancing operation completes, AIX iSCSI initiators lose volume connectivity. This

issue can be corrected by applying a patch to your AIX system. For more information, contact your IBM

customer support provider and reference incident number IZ90885, or visit

ibm.com/support/docview.wss?uid=isg1IZ90885&crawler=1.

Note: As a short-term workaround, before applying the patch, disable the PS Series group's load-balancing

feature.

Page is loading ...

Dell Storage Solution Resources Owner's manual

Brand: Dell

Model: Storage Solution Resources

Type: Owner's manual

With Dell Storage Solution Resources, you can:

Centralize data storage: Consolidate data from various sources onto a single, unified platform, simplifying data management and enhancing accessibility.
Ensure data protection: Implement comprehensive data protection strategies with features like snapshots, replication, and backup, ensuring business continuity and minimizing data loss risks.

Download PDF

report

Dell Storage Solution Resources Owner's manual

Dell Storage Solution Resources Owner's manual

Related papers

Other documents